The answer to your specific question "Is this correct?" is "yes, I believe so".

As to the statement "Once a degree of freedom has been fixed in one displacement vector, it cannot be free in another displacement vector at the same node (leaving a displacement field blank will default to zero in this case)", think of it this way:

Say you have a double acting "Y" (vertical) support that for some reason grows up by 0.5" in the operating condition. There are several points to recognize here:

1) Numerically speaking, a restraint is displacment with a magnitude of zero.

2) If you define a boundary condition (restraint or displacment) at a node point, that boundary condition exists regardless of the load case in question. (You don't run out to the field and add or remove a restraint before changing temperatures or pressures.)

So, if in Displacment Vector 1, you define DY=0.5, then you have fixed "DY" for all nine displacement vectors. If you don't specify a numeric value, CAESAR II will assume 0.0. You can't say that for this specific node, it has a displacement in one vector but is free in another vector. (That would mean you ran out to the field and changed the system.)

The same assumption is made when you analyze the Sustained Load Case. The load primatives making up the case would be W + P1, note the absence of D1. Because you defined "DY=0.5", but did not include "D1" in the load case, CAESAR II will assume 0.0 in the DY direction at this node point for the Sustained case. (Again because the boundary condition must be consistent.)